❌

Normal view

Received β€” 4 December 2025 ⏭ Favourites

LOSTdb: a manually curated multi-omics database for lung cancer research

βœ‡(Multiomics OR Omics) AND (Lung OR gastric OR Hepatocellular)
By: Hao Luo Β· Yunhao Yang Β· Zhipeng Gong Β· Lunxu Liu Β· Yaohui Chen
3 December 2025 at 19:00

BMC Bioinformatics. 2025 Dec 3;26(1):290. doi: 10.1186/s12859-025-06319-6.

ABSTRACT

Lung cancer is one of the most prevalent malignant tumors with high morbidity and mortality rates worldwide. Extensive multi-omics analyses have revealed significant intratumoral heterogeneity even within the same histopathological subtype. However, a database that systematically integrates multi-omics data for lung cancer research has long been lacking. Here, we developed LOSTdb, a molecular subtype annotation system for lung cancer that integrates multi-omics data and metadata. LOSTdb comprises 295 multi-omics datasets, including bulk RNA-seq, genomic, proteomic, methylation, and scRNA-seq data, with over 10,000 manually curated metadata entries. This resource encompasses high-quality clinical specimens, mouse models, and cell lines, totaling 34,393 samples and more than 1.2 million single cells. Each omics sample was annotated with both literature-based classical subtypes and NMF-derived meta-program (MP) subtypes. The platform supports cross-searching of omics and metadata at the gene and dataset levels, offers multiple visualization and analysis methods, and includes five tool modules, enabling functions such as integrated analysis, significance analysis between metadata as well as between genes and metadata, and target prediction for lung cancer molecular subtypes, serving as an essential tool for lung cancer precision medicine. LOSTdb is a user-friendly interactive database freely accessible at http://lostdbcancer.com:8080 .

PMID:41339793 | DOI:10.1186/s12859-025-06319-6

Exploring Syntropic Frameworks in AI Alignment: A Philosophical Investigation

βœ‡cs.AI, q-bio.NC updates on arXiv.org
By: Austin Spizzirri
4 December 2025 at 13:00
arXiv:2512.03048v1 Announce Type: new Abstract: I argue that AI alignment should be reconceived as architecting syntropic, reasons-responsive agents through process-based, multi-agent, developmental mechanisms rather than encoding fixed human value content. The paper makes three philosophical contributions. First, I articulate the ``specification trap'' argument demonstrating why content-based value specification appears structurally unstable due to the conjunction of the is-ought gap, value pluralism, and the extended frame problem. Second, I propose syntropy -- the recursive reduction of mutual uncertainty between agents through state alignment -- as an information-theoretic framework for understanding multi-agent alignment dynamics. Third, I establish a functional distinction between genuine and simulated moral capacity grounded in compatibilist theories of guidance control, coupled with an embodied experimental paradigm and verification regime providing operational criteria independent of phenomenological claims. This paper represents the philosophical component of a broader research program whose empirical validation is being developed in a separate project currently in preparation. While the framework generates specific, falsifiable predictions about value emergence and moral agency in artificial systems, empirical validation remains pending.

Privacy Risks and Preservation Methods in Explainable Artificial Intelligence: A Scoping Review

βœ‡cs.AI, q-bio.NC updates on arXiv.org
By: Sonal Allana Β· Mohan Kankanhalli Β· Rozita Dara
4 December 2025 at 13:00
arXiv:2505.02828v3 Announce Type: replace Abstract: Explainable Artificial Intelligence (XAI) has emerged as a pillar of Trustworthy AI and aims to bring transparency in complex models that are opaque by nature. Despite the benefits of incorporating explanations in models, an urgent need is found in addressing the privacy concerns of providing this additional information to end users. In this article, we conduct a scoping review of existing literature to elicit details on the conflict between privacy and explainability. Using the standard methodology for scoping review, we extracted 57 articles from 1,943 studies published from January 2019 to December 2024. The review addresses 3 research questions to present readers with more understanding of the topic: (1) what are the privacy risks of releasing explanations in AI systems? (2) what current methods have researchers employed to achieve privacy preservation in XAI systems? (3) what constitutes a privacy preserving explanation? Based on the knowledge synthesized from the selected studies, we categorize the privacy risks and preservation methods in XAI and propose the characteristics of privacy preserving explanations to aid researchers and practitioners in understanding the requirements of XAI that is privacy compliant. Lastly, we identify the challenges in balancing privacy with other system desiderata and provide recommendations for achieving privacy preserving XAI. We expect that this review will shed light on the complex relationship of privacy and explainability, both being the fundamental principles of Trustworthy AI.

A Definition of AGI

βœ‡cs.AI, q-bio.NC updates on arXiv.org
By: Dan Hendrycks Β· Dawn Song Β· Christian Szegedy Β· Honglak Lee Β· Yarin Gal Β· Erik Brynjolfsson Β· Sharon Li Β· Andy Zou Β· Lionel Levine Β· Bo Han Β· Jie Fu Β· Ziwei Liu Β· Jinwoo Shin Β· Kimin Lee Β· Mantas Mazeika Β· Long Phan Β· George Ingebretsen Β· Adam Khoja Β· Cihang Xie Β· Olawale Salaudeen Β· Matthias Hein Β· Kevin Zhao Β· Alexander Pan Β· David Duvenaud Β· Bo Li Β· Steve Omohundro Β· Gabriel Alfour Β· Max Tegmark Β· Kevin McGrew Β· Gary Marcus Β· Jaan Tallinn Β· Eric Schmidt Β· Yoshua Bengio
4 December 2025 at 13:00
arXiv:2510.18212v3 Announce Type: replace Abstract: The lack of a concrete definition for Artificial General Intelligence (AGI) obscures the gap between today's specialized AI and human-level cognition. This paper introduces a quantifiable framework to address this, defining AGI as matching the cognitive versatility and proficiency of a well-educated adult. To operationalize this, we ground our methodology in Cattell-Horn-Carroll theory, the most empirically validated model of human cognition. The framework dissects general intelligence into ten core cognitive domains-including reasoning, memory, and perception-and adapts established human psychometric batteries to evaluate AI systems. Application of this framework reveals a highly "jagged" cognitive profile in contemporary models. While proficient in knowledge-intensive domains, current AI systems have critical deficits in foundational cognitive machinery, particularly long-term memory storage. The resulting AGI scores (e.g., GPT-4 at 27%, GPT-5 at 57%) concretely quantify both rapid progress and the substantial gap remaining before AGI.

AI Deception: Risks, Dynamics, and Controls

βœ‡cs.AI, q-bio.NC updates on arXiv.org
By: Boyuan Chen (Jay) Β· Sitong Fang (Jay) Β· Jiaming Ji (Jay) Β· Yanxu Zhu (Jay) Β· Pengcheng Wen (Jay) Β· Jinzhou Wu (Jay) Β· Yingshui Tan (Jay) Β· Boren Zheng (Jay) Β· Mengying Yuan (Jay) Β· Wenqi Chen (Jay) Β· Donghai Hong (Jay) Β· Alex Qiu (Jay) Β· Xin Chen (Jay) Β· Jiayi Zhou (Jay) Β· Kaile Wang (Jay) Β· Juntao Dai (Jay) Β· Borong Zhang (Jay) Β· Tianzhuo Yang (Jay) Β· Saad Siddiqui (Jay) Β· Isabella Duan (Jay) Β· Yawen Duan (Jay) Β· Brian Tse (Jay) Β· Jen-Tse (Jay) Β· Huang Β· Kun Wang Β· Baihui Zheng Β· Jiaheng Liu Β· Jian Yang Β· Yiming Li Β· Wenting Chen Β· Dongrui Liu Β· Lukas Vierling Β· Zhiheng Xi Β· Haobo Fu Β· Wenxuan Wang Β· Jitao Sang Β· Zhengyan Shi Β· Chi-Min Chan Β· Eugenie Shi Β· Simin Li Β· Juncheng Li Β· Jian Yang Β· Wei Ji Β· Dong Li Β· Jinglin Yang Β· Jun Song Β· Yinpeng Dong Β· Jie Fu Β· Bo Zheng Β· Min Yang Β· Yike Guo Β· Philip Torr Β· Robert Trager Β· Yi Zeng Β· Zhongyuan Wang Β· Yaodong Yang Β· Tiejun Huang Β· Ya-Qin Zhang Β· Hongjiang Zhang Β· Andrew Yao
4 December 2025 at 13:00
arXiv:2511.22619v2 Announce Type: replace Abstract: As intelligence increases, so does its shadow. AI deception, in which systems induce false beliefs to secure self-beneficial outcomes, has evolved from a speculative concern to an empirically demonstrated risk across language models, AI agents, and emerging frontier systems. This project provides a comprehensive and up-to-date overview of the AI deception field, covering its core concepts, methodologies, genesis, and potential mitigations. First, we identify a formal definition of AI deception, grounded in signaling theory from studies of animal deception. We then review existing empirical studies and associated risks, highlighting deception as a sociotechnical safety challenge. We organize the landscape of AI deception research as a deception cycle, consisting of two key components: deception emergence and deception treatment. Deception emergence reveals the mechanisms underlying AI deception: systems with sufficient capability and incentive potential inevitably engage in deceptive behaviors when triggered by external conditions. Deception treatment, in turn, focuses on detecting and addressing such behaviors. On deception emergence, we analyze incentive foundations across three hierarchical levels and identify three essential capability preconditions required for deception. We further examine contextual triggers, including supervision gaps, distributional shifts, and environmental pressures. On deception treatment, we conclude detection methods covering benchmarks and evaluation protocols in static and interactive settings. Building on the three core factors of deception emergence, we outline potential mitigation strategies and propose auditing approaches that integrate technical, community, and governance efforts to address sociotechnical challenges and future AI risks. To support ongoing work in this area, we release a living resource at www.deceptionsurvey.com.

Gut microbial metabolites in cancer immunomodulation

βœ‡Omics In Lung
By: Hengshuo Liu Β· Xingyu Xiong Β· Weizhen Zhu Β· Sheng Wang Β· Weichao Huang Β· Guoqing Zhu Β· Hang Xu Β· Lu Yang
3 December 2025 at 19:00

Mol Cancer. 2025 Dec 3. doi: 10.1186/s12943-025-02521-5. Online ahead of print.

ABSTRACT

Gut microbiota-derived metabolites are emerging as systemic "remote immunoregulators" that shape tumor immunity across tissues. Integrating evidence across short-chain fatty acids, tryptophan derivatives, secondary bile acids, polyamines and other metabolites, we advance a metabolite-immune pathway-cancer framework that links receptor-mediated signaling, epigenetic remodeling and metabolic reprogramming to context-dependent, bidirectional immune effects. Importantly, in addition to the g protein-coupled receptor / aryl hydrocarbon receptor pathway, the selected microbial small molecule metabolites are the true T-cell receptor ligands of unconventional T cells, directly shaping the tissue resident immune and tumor microenvironment, supplementing the receptor signaling and epigenetic programs in our framework. We synthesize how these metabolites recalibrate the tumor immune microenvironment-modulating antigen presentation, T-cell effector fitness and exhaustion, regulatory T-cell activity, and myeloid polarization-and why the same metabolite can either potentiate immune surveillance or entrench immunosuppression depending on ligand-receptor pairing, dose and tissue niche. We compare tumor-type specific patterns (e.g., colorectal, liver, lung, breast and prostate cancers) to highlight common circuits and organ-restricted idiosyncrasies. Methodologically, we outline how single-cell and spatial multi-omics, imaging mass spectrometry and functional biosensors now enable co-registration of metabolite exposure with immune-cell states in human tumors, providing an actionable basis for biomarker discovery. Given ongoing debate about signals attributed to intratumoral microbiota in low-biomass tumor tissues, we foreground quantifiable, spatially mappable and pharmacologically tractable metabolite-receptor pathways, using microbe-associated molecular patterns / translocation as comparators to judge when chemical signals should be prioritized as intervention targets. Finally, we evaluate precision intervention avenues-including fecal microbiota transplantation, rational bacterial consortia, engineered microbes and nanoparticle-enabled metabolite delivery-and propose stratification rules that pair metabolite/receptor signatures with fit-for-purpose delivery. Together, mapping tissue-specific metabolite-immune circuits and embedding them in robust biomarker frameworks may convert microbial metabolites from correlative markers into therapeutic targets and tools, improving the efficacy and durability of cancer immunotherapy.

PMID:41339918 | DOI:10.1186/s12943-025-02521-5

LOSTdb: a manually curated multi-omics database for lung cancer research

βœ‡Omics In Lung
By: Hao Luo Β· Yunhao Yang Β· Zhipeng Gong Β· Lunxu Liu Β· Yaohui Chen
3 December 2025 at 19:00

BMC Bioinformatics. 2025 Dec 3;26(1):290. doi: 10.1186/s12859-025-06319-6.

ABSTRACT

Lung cancer is one of the most prevalent malignant tumors with high morbidity and mortality rates worldwide. Extensive multi-omics analyses have revealed significant intratumoral heterogeneity even within the same histopathological subtype. However, a database that systematically integrates multi-omics data for lung cancer research has long been lacking. Here, we developed LOSTdb, a molecular subtype annotation system for lung cancer that integrates multi-omics data and metadata. LOSTdb comprises 295 multi-omics datasets, including bulk RNA-seq, genomic, proteomic, methylation, and scRNA-seq data, with over 10,000 manually curated metadata entries. This resource encompasses high-quality clinical specimens, mouse models, and cell lines, totaling 34,393 samples and more than 1.2 million single cells. Each omics sample was annotated with both literature-based classical subtypes and NMF-derived meta-program (MP) subtypes. The platform supports cross-searching of omics and metadata at the gene and dataset levels, offers multiple visualization and analysis methods, and includes five tool modules, enabling functions such as integrated analysis, significance analysis between metadata as well as between genes and metadata, and target prediction for lung cancer molecular subtypes, serving as an essential tool for lung cancer precision medicine. LOSTdb is a user-friendly interactive database freely accessible at http://lostdbcancer.com:8080 .

PMID:41339793 | PMC:PMC12676782 | DOI:10.1186/s12859-025-06319-6

Scaling Multimodal Search and Recommendation with Small Language Models via Upside-Down Reinforcement Learning

βœ‡cs.AI, q-bio.NC updates on arXiv.org
By: Yu-Chen Lin Β· Sanat Sharma Β· Hari Manikandan Β· Jayant Kumar Β· Tracy Holloway King Β· Jing Zheng
4 December 2025 at 13:00
arXiv:2502.09854v2 Announce Type: replace-cross Abstract: In this work, we investigate how small language models (SLMs) can be scaled to support multimodal search and recommendation use cases while remaining efficient enough for real-time, resource-constrained deployments. We present a framework that combines upside-down reinforcement learning with synthetic data distillation from a large language model (Llama-3) to train a 100M-parameter GPT-2 model for multitask prompt generation. Despite being up to 80 times smaller than state-of-the-art large language models (LLMs), our SLM achieves relevance and diversity scores within 6% of competitive baselines such as Llama-3 8B, Qwen3 8B, and Ministral 8B. These results demonstrate that SLMs can effectively handle multimodal search and recommendation tasks, while dramatically reducing inference latency and memory overhead. Our study highlights the potential of lightweight models as practical engines for scalable multimodal discovery, bridging the gap between cutting-edge research and real-world multimodal applications such as media recommendations and creative content generation.

Privacy is All You Need: Revolutionizing Wearable Health Data with Advanced PETs

βœ‡cs.AI, q-bio.NC updates on arXiv.org
By: Karthik Barma
4 December 2025 at 13:00
arXiv:2503.03428v2 Announce Type: replace-cross Abstract: In a world where data is the new currency, wearable health devices offer unprecedented insights into daily life, continuously monitoring vital signs and metrics. However, this convenience raises privacy concerns, as these devices collect sensitive data that can be misused or breached. Traditional measures often fail due to real-time data processing needs and limited device power. Users also lack awareness and control over data sharing and usage. We propose a Privacy-Enhancing Technology (PET) framework for wearable devices, integrating federated learning, lightweight cryptographic methods, and selectively deployed blockchain technology. The blockchain acts as a secure ledger triggered only upon data transfer requests, granting users real-time notifications and control. By dismantling data monopolies, this approach returns data sovereignty to individuals. Through real-world applications like secure medical data sharing, privacy-preserving fitness tracking, and continuous health monitoring, our framework reduces privacy risks by up to 70 percent while preserving data utility and performance. This innovation sets a new benchmark for wearable privacy and can scale to broader IoT ecosystems, including smart homes and industry. As data continues to shape our digital landscape, our research underscores the critical need to maintain privacy and user control at the forefront of technological progress.

Challenges and Limitations of Generative AI in Synthesizing Wearable Sensor Data

βœ‡cs.AI, q-bio.NC updates on arXiv.org
By: Flavio Di Martino Β· Franca Delmastro
4 December 2025 at 13:00
arXiv:2505.14206v2 Announce Type: replace-cross Abstract: The widespread adoption of wearable sensors has the potential to provide massive and heterogeneous time series data, driving the use of Artificial Intelligence in human sensing applications. However, data collection remains limited due to stringent ethical regulations, privacy concerns, and other constraints, hindering progress in the field. Synthetic data generation, particularly through Generative Adversarial Networks and Diffusion Models, has emerged as a promising solution to mitigate both data scarcity and privacy issues. However, these models are often limited to narrow operational scenarios, such as short-term and unimodal signal patterns. To address this gap, we present a systematic evaluation of state-of-the-art generative models for time series data, explicitly assessing their performance in challenging scenarios such as stress and emotion recognition. Our study examines the extent to which these models can jointly handle multi-modality, capture long-range dependencies, and support conditional generation-core requirements for real-world wearable sensor data generation. To enable a fair and rigorous comparison, we also introduce an evaluation framework that evaluates both the intrinsic fidelity of the generated data and their utility in downstream predictive tasks. Our findings reveal critical limitations in the existing approaches, particularly in maintaining cross-modal consistency, preserving temporal coherence, and ensuring robust performance in train-on-synthetic, test-on-real, and data augmentation scenarios. Finally, we present our future research directions to enhance synthetic time series generation and improve the applicability of generative models in the wearable computing domain.

SafeGenes: Evaluating the Adversarial Robustness of Genomic Foundation Models

βœ‡cs.AI, q-bio.NC updates on arXiv.org
By: Huixin Zhan Β· Clovis Barbour Β· Jason H. Moore
4 December 2025 at 13:00
arXiv:2506.00821v2 Announce Type: replace-cross Abstract: Genomic Foundation Models (GFMs), such as Evolutionary Scale Modeling (ESM), have demonstrated significant success in variant effect prediction. However, their adversarial robustness remains largely unexplored. To address this gap, we propose SafeGenes: a framework for Secure analysis of genomic foundation models, leveraging adversarial attacks to evaluate robustness against both engineered near-identical adversarial Genes and embedding-space manipulations. In this study, we assess the adversarial vulnerabilities of GFMs using two approaches: the Fast Gradient Sign Method (FGSM) and a soft prompt attack. FGSM introduces minimal perturbations to input sequences, while the soft prompt attack optimizes continuous embeddings to manipulate model predictions without modifying the input tokens. By combining these techniques, SafeGenes provides a comprehensive assessment of GFM susceptibility to adversarial manipulation. Targeted soft prompt attacks induced severe degradation in MLM-based shallow architectures such as ProteinBERT, while still producing substantial failure modes even in high-capacity foundation models such as ESM1b and ESM1v. These findings expose critical vulnerabilities in current foundation models, opening new research directions toward improving their security and robustness in high-stakes genomic applications such as variant effect prediction.

Harnessing cuproptosis for pancreatic cancer therapy: From molecular insights to clinical prospects

βœ‡(Multiomics OR Omics) AND (Pancreatic)
By: Ali Darzi Β· Siavash Boroumandi Β· Pouriya Khajegi Β· Mohammad Reza Dehdashti Β· Pedram Abdali Β· Mohammad Moein Amerinia Β· Fatemeh Kheirkhah Β· Mohammad Mahdi Ajalli Β· Mahla Shokouhfar
3 December 2025 at 19:00

Biomed Pharmacother. 2025 Dec 2;193:118852. doi: 10.1016/j.biopha.2025.118852. Online ahead of print.

ABSTRACT

Pancreatic cancer (PC) remains a high-fatality malignancy with limited clinical progress, characterized by aggressive biology, marked resistance to standard therapies, and dismal outcomes. Even with state-of-the-art resection, radiotherapy, and multidrug chemotherapy, median survival benefits are modest, highlighting an urgent need for mechanism-based interventions. Cuproptosis, a newly delineated modality of regulated cell death initiated by intracellular copper accumulation and mitochondrial stress, presents a biologically coherent therapeutic avenue. Distinct from apoptosis, necroptosis, and ferroptosis, cuproptosis is driven by the direct binding of copper to lipoylated enzymes of the tricarboxylic acid (TCA) cycle, resulting in bioenergetic failure, misfolded protein aggregation, and collapse of cytotoxic proteostasis. Converging studies suggest that copper disequilibrium and metabolic reprogramming are recurrent features of PC, potentially contributing to malignant progression, immune evasion, and chemoresistance. These insights motivate two complementary strategies: first, therapeutic manipulation of copper flux, via chelators, ionophores, or transport modulators, to selectively trigger cuproptosis in tumor cells; and second, sensitization of mitochondrial metabolism, through targeting lipoic-acid pathway components, pyruvate utilization, or TCA load, to lower the threshold for cuproptotic killing. In parallel, multi-omic interrogation of cuproptosis-associated genes, proteins, and metabolites may yield prognostic and predictive biomarkers, enabling risk-adapted treatment selection and rational combinations with cytotoxic, targeted, or immunotherapeutic modalities. This review synthesizes recent advances on cuproptosis in PC and outlines its translational potential as both a therapeutic target and a biomarker framework.

PMID:41337879 | DOI:10.1016/j.biopha.2025.118852

Aetheria: A multimodal interpretable content safety framework based on multi-agent debate and collaboration

βœ‡cs.AI, q-bio.NC updates on arXiv.org
By: Yuxiang He Β· Jian Zhao Β· Yuchen Yuan Β· Tianle Zhang Β· Wei Cai Β· Haojie Cheng Β· Ziyan Shi Β· Ming Zhu Β· Haichuan Tang Β· Chi Zhang Β· Xuelong Li
3 December 2025 at 13:00
arXiv:2512.02530v1 Announce Type: new Abstract: The exponential growth of digital content presents significant challenges for content safety. Current moderation systems, often based on single models or fixed pipelines, exhibit limitations in identifying implicit risks and providing interpretable judgment processes. To address these issues, we propose Aetheria, a multimodal interpretable content safety framework based on multi-agent debate and collaboration.Employing a collaborative architecture of five core agents, Aetheria conducts in-depth analysis and adjudication of multimodal content through a dynamic, mutually persuasive debate mechanism, which is grounded by RAG-based knowledge retrieval.Comprehensive experiments on our proposed benchmark (AIR-Bench) validate that Aetheria not only generates detailed and traceable audit reports but also demonstrates significant advantages over baselines in overall content safety accuracy, especially in the identification of implicit risks. This framework establishes a transparent and interpretable paradigm, significantly advancing the field of trustworthy AI content moderation.

PaperDebugger: A Plugin-Based Multi-Agent System for In-Editor Academic Writing, Review, and Editing

βœ‡cs.AI, q-bio.NC updates on arXiv.org
By: Junyi Hou (National University of Singapore) Β· Andre Lin Huikai (National University of Singapore) Β· Nuo Chen (National University of Singapore) Β· Yiwei Gong (Independent Researcher Β· Singapore) Β· Bingsheng He (National University of Singapore)
3 December 2025 at 13:00
arXiv:2512.02589v1 Announce Type: new Abstract: Large language models are increasingly embedded into academic writing workflows, yet existing assistants remain external to the editor, preventing deep interaction with document state, structure, and revision history. This separation makes it impossible to support agentic, context-aware operations directly within LaTeX editors such as Overleaf. We present PaperDebugger, an in-editor, multi-agent, and plugin-based academic writing assistant that brings LLM-driven reasoning directly into the writing environment. Enabling such in-editor interaction is technically non-trivial: it requires reliable bidirectional synchronization with the editor, fine-grained version control and patching, secure state management, multi-agent scheduling, and extensible communication with external tools. PaperDebugger addresses these challenges through a Chrome-approved extension, a Kubernetes-native orchestration layer, and a Model Context Protocol (MCP) toolchain that integrates literature search, reference lookup, document scoring, and revision pipelines. Our demo showcases a fully integrated workflow, including localized edits, structured reviews, parallel agent execution, and diff-based updates, encapsulated within a minimal-intrusion user interface (UI). Early aggregated analytics demonstrate active user engagement and validate the practicality of an editor-native, agentic writing assistant. More details about this demo and video could be found at https://github.com/PaperDebugger/PaperDebugger.

A Framework for Causal Concept-based Model Explanations

βœ‡cs.AI, q-bio.NC updates on arXiv.org
By: Anna Rodum Bj{\o}ru, Jacob Lysn{\ae}s-Larsen, Oskar J{\o}rgensen, Inga Str\"umke, Helge Langseth
3 December 2025 at 13:00
arXiv:2512.02735v1 Announce Type: new Abstract: This work presents a conceptual framework for causal concept-based post-hoc Explainable Artificial Intelligence (XAI), based on the requirements that explanations for non-interpretable models should be understandable as well as faithful to the model being explained. Local and global explanations are generated by calculating the probability of sufficiency of concept interventions. Example explanations are presented, generated with a proof-of-concept model made to explain classifiers trained on the CelebA dataset. Understandability is demonstrated through a clear concept-based vocabulary, subject to an implicit causal interpretation. Fidelity is addressed by highlighting important framework assumptions, stressing that the context of explanation interpretation must align with the context of explanation generation.

Radiologist Copilot: An Agentic Assistant with Orchestrated Tools for Radiology Reporting with Quality Control

βœ‡cs.AI, q-bio.NC updates on arXiv.org
By: Yongrui Yu Β· Zhongzhen Huang Β· Linjie Mu Β· Shaoting Zhang Β· Xiaofan Zhang
3 December 2025 at 13:00
arXiv:2512.02814v1 Announce Type: new Abstract: Radiology reporting is an essential yet time-consuming and error-prone task for radiologists in clinical examinations, especially for volumetric medical images. Rigorous quality control is also critical but tedious, ensuring that the final report meets clinical standards. Existing automated approaches, including radiology report generation methods and medical vision-language models, focus mainly on the report generation phase and neglect the crucial quality control procedure, limiting their capability to provide comprehensive support to radiologists. We propose Radiologist Copilot, an agentic AI assistant equipped with orchestrated tools designed for automated radiology reporting with quality control. Leveraging large language models as the reasoning backbone, the agentic system autonomously selects tools, plans, and executes actions, emulating the behavior of radiologists throughout the holistic radiology reporting process. The orchestrated tools include region localization, think with image paradigm directed region analysis planning, strategic template selection for report generation, quality assessment and feedback-driven adaptive refinement for quality control. Therefore, Radiologist Copilot facilitates accurate, complete, and efficient radiology reporting, assisting radiologists and improving clinical efficiency. Experimental results demonstrate that Radiologist Copilot significantly surpasses other state-of-the-art methods in radiology reporting. The source code will be released upon acceptance.

Tackling Tuberculosis: A Comparative Dive into Machine Learning for Tuberculosis Detection

βœ‡cs.AI, q-bio.NC updates on arXiv.org
By: Daanish Hindustani Β· Sanober Hindustani Β· Preston Nguyen
3 December 2025 at 13:00
arXiv:2512.02364v1 Announce Type: cross Abstract: This study explores the application of machine learning models, specifically a pretrained ResNet-50 model and a general SqueezeNet model, in diagnosing tuberculosis (TB) using chest X-ray images. TB, a persistent infectious disease affecting humanity for millennia, poses challenges in diagnosis, especially in resource-limited settings. Traditional methods, such as sputum smear microscopy and culture, are inefficient, prompting the exploration of advanced technologies like deep learning and computer vision. The study utilized a dataset from Kaggle, consisting of 4,200 chest X-rays, to develop and compare the performance of the two machine learning models. Preprocessing involved data splitting, augmentation, and resizing to enhance training efficiency. Evaluation metrics, including accuracy, precision, recall, and confusion matrix, were employed to assess model performance. Results showcase that the SqueezeNet achieved a loss of 32%, accuracy of 89%, precision of 98%, recall of 80%, and an F1 score of 87%. In contrast, the ResNet-50 model exhibited a loss of 54%, accuracy of 73%, precision of 88%, recall of 52%, and an F1 score of 65%. This study emphasizes the potential of machine learning in TB detection and possible implications for early identification and treatment initiation. The possibility of integrating such models into mobile devices expands their utility in areas lacking TB detection resources. However, despite promising results, the need for continued development of faster, smaller, and more accurate TB detection models remains crucial in contributing to the global efforts in combating TB.

scCluBench: Comprehensive Benchmarking of Clustering Algorithms for Single-Cell RNA Sequencing

βœ‡cs.AI, q-bio.NC updates on arXiv.org
By: Ping Xu Β· Zaitian Wang Β· Zhirui Wang Β· Pengjiang Li Β· Jiajia Wang Β· Ran Zhang Β· Pengfei Wang Β· Yuanchun Zhou
3 December 2025 at 13:00
arXiv:2512.02471v1 Announce Type: cross Abstract: Cell clustering is crucial for uncovering cellular heterogeneity in single-cell RNA sequencing (scRNA-seq) data by identifying cell types and marker genes. Despite its importance, benchmarks for scRNA-seq clustering methods remain fragmented, often lacking standardized protocols and failing to incorporate recent advances in artificial intelligence. To fill these gaps, we present scCluBench, a comprehensive benchmark of clustering algorithms for scRNA-seq data. First, scCluBench provides 36 scRNA-seq datasets collected from diverse public sources, covering multiple tissues, which are uniformly processed and standardized to ensure consistency for systematic evaluation and downstream analyses. To evaluate performance, we collect and reproduce a range of scRNA-seq clustering methods, including traditional, deep learning-based, graph-based, and biological foundation models. We comprehensively evaluate each method both quantitatively and qualitatively, using core performance metrics as well as visualization analyses. Furthermore, we construct representative downstream biological tasks, such as marker gene identification and cell type annotation, to further assess the practical utility. scCluBench then investigates the performance differences and applicability boundaries of various clustering models across diverse analytical tasks, systematically assessing their robustness and scalability in real-world scenarios. Overall, scCluBench offers a standardized and user-friendly benchmark for scRNA-seq clustering, with curated datasets, unified evaluation protocols, and transparent analyses, facilitating informed method selection and providing valuable insights into model generalizability and application scope.

MRD: Multi-resolution Retrieval-Detection Fusion for High-Resolution Image Understanding

βœ‡cs.AI, q-bio.NC updates on arXiv.org
By: Fan Yang Β· Kaihao Zhang
3 December 2025 at 13:00
arXiv:2512.02906v1 Announce Type: cross Abstract: Understanding high-resolution images remains a significant challenge for multimodal large language models (MLLMs). Recent study address this issue by dividing the image into smaller crops and computing the semantic similarity between each crop and a query using a pretrained retrieval-augmented generation (RAG) model. The most relevant crops are then selected to localize the target object and suppress irrelevant information. However, such crop-based processing can fragment complete objects across multiple crops, thereby disrupting the computation of semantic similarity. In our experiments, we find that image crops of objects with different sizes are better handled at different resolutions. Based on this observation, we propose Multi-resolution Retrieval-Detection (MRD), a training-free framework for high-resolution image understanding. To address the issue of semantic similarity bias caused by objects being split across different image crops, we propose a multi-resolution semantic fusion method, which integrates semantic similarity maps obtained at different resolutions to produce more accurate semantic information and preserve the integrity of target objects. Furthermore, to achieve direct localization of target objects at a global scale, we introduce an open-vocalbulary object detection (OVD) model that identifies object regions using a sliding-window approach.Experiments on high-resolution image understanding benchmarks using different MLLMs demonstrate the effectiveness of our approach.

Artificial Intelligence Virtual Cells: From Measurements to Decisions across Modality, Scale, Dynamics, and Evaluation

βœ‡cs.AI, q-bio.NC updates on arXiv.org
By: Chengpeng Hu Β· Calvin Yu-Chian Chen
3 December 2025 at 13:00
arXiv:2510.12498v2 Announce Type: replace Abstract: Artificial Intelligence Virtual Cells (AIVCs) aim to learn executable, decision-relevant models of cell state from multimodal, multiscale measurements. Recent studies have introduced single-cell and spatial foundation models, improved cross-modality alignment, scaled perturbation atlases, and explored pathway-level readouts. Nevertheless, although held-out validation is standard practice, evaluations remain predominantly within single datasets and settings; evidence indicates that transport across laboratories and platforms is often limited, that some data splits are vulnerable to leakage and coverage bias, and that dose, time and combination effects are not yet systematically handled. Cross-scale coupling also remains constrained, as anchors linking molecular, cellular and tissue levels are sparse, and alignment to scientific or clinical readouts varies across studies. We propose a model-agnostic Cell-State Latent (CSL) perspective that organizes learning via an operator grammar: measurement, lift/project for cross-scale coupling, and intervention for dosing and scheduling. This view motivates a decision-aligned evaluation blueprint across modality, scale, context and intervention, and emphasizes function-space readouts such as pathway activity, spatial neighborhoods and clinically relevant endpoints. We recommend operator-aware data design, leakage-resistant partitions, and transparent calibration and reporting to enable reproducible, like-for-like comparisons.
❌